def create_arrow_map(events: pd.DataFrame,
event_type: str = "Pass",
match: str = "an unknown match",
fig=None,
ax=None,
pitch_length: int = 120,
pitch_width: int = 80,
color: str = "blue"):
"""
This functions creates a map of arrows between the start point and end point of an event.
:param events: a data frame containing events
:param event_type: a string describing the event type for which to plot arrows
:param match: a string describing the match for which the events are plotted
:param fig: a figure
:param ax: a subplot of the figure on which to plot the arrows
:param pitch_length: a integer describing the length of the pitch
:param pitch_width: a integer describing the width of the pitch
:param color: a string describing the color of the arrows
:return fig, ax: a figure containing a pitch with event arrows
"""
# Filter only relevant events
temp_df = events[events.type_name == event_type]
# Filter start and end locations
locations = pd.DataFrame(temp_df.location.to_list(),
columns=['x_start', 'y_start'])
locations[['x_end', 'y_end']] = pd.DataFrame(
temp_df[f'{event_type.lower()}_end_location'].to_list())
# Add delta_x and delta_y
locations['dx'] = locations.x_end - locations.x_start
locations['dy'] = locations.y_end - locations.y_start
# Draw pitch
fig, ax = create_pitch(pitch_length, pitch_width, fig, ax)
# Draw arrows
for i, event in locations.iterrows():
x_start, y_start, x_end, y_end, dx, dy = list(event)
pass_circle = plt.Circle((x_start, y_start), 2, color=color)
pass_circle.set_alpha(0.2)
pass_arrow = plt.Arrow(x_start,
y_start,
dx,
dy,
width=3,
color=color,
in_layout=True)
ax.add_patch(pass_circle)
ax.add_patch(pass_arrow)
# Set limits
plt.ylim(0, pitch_width)
plt.xlim(0, pitch_length)
plt.title(
f"{event_type} map of {events.player_name.unique()[0]} during {match}")
return fig, ax
def create_event_map(data: pd.DataFrame, fig=None, ax=None, pitch_length: int = 120, pitch_width: int = 80):
# Draw pitch
fig, ax = create_pitch(pitch_length, pitch_width, fig, ax)
# For all events get type, start and end location, and determine outcome
locations = pd.DataFrame(data.apply(lambda x: map_location(x), axis=1).to_list(), columns=['x_start', 'y_start'])
locations['type_name'] = pd.DataFrame(data['type_name'].to_list())
locations[['x_end', 'y_end', 'z_end']] = pd.DataFrame(data.apply(lambda x: map_end_location(x), axis=1).to_list())
locations[['dx', 'dy']] = pd.DataFrame(locations.apply(lambda x: map_dx_dy(x), axis=1).to_list())
locations['outcome'] = pd.DataFrame(data.apply(lambda x: map_outcome(x), axis=1).to_list())
# Plot
for event_type in locations.type_name.unique():
temp = locations[locations['type_name'] == event_type]
if event_type == 'Ball Receipt*':
temp.plot(kind='scatter', x='x_start', y='y_start', marker='o', s=50, color='tab:cyan',
label='Ball Receipt', alpha=1, ax=ax, zorder=2)
if event_type == 'Carry':
temp.plot(kind='scatter', x='x_start', y='y_start', marker='p', s=40, color='blue',
label='Carry', alpha=1, ax=ax, zorder=0)
for i, event in temp.iterrows():
ax.arrow(event.x_start, event.y_start, event.dx, event.dy, head_width=0.5, ls=':', color='blue',
alpha=1, zorder=0, length_includes_head=True)
if event_type == 'Dribble':
complete = temp[temp['outcome'] == 'Complete']
incomplete = temp[temp['outcome'] == 'Incomplete']
complete.plot(kind='scatter', x='x_start', y='y_start', marker='X', s=75, color='tab:green',
label='Complete Dribble', alpha=1, ax=ax, zorder=3, lw=0.5, edgecolor='black')
incomplete.plot(kind='scatter', x='x_start', y='y_start', marker='X', s=75, color='tab:red',
label='Incomplete Dribble', alpha=1, ax=ax, zorder=3, lw=0.5, edgecolor='black')
if event_type == 'Pass':
complete = temp[temp['outcome'] != 'Incomplete']
incomplete = temp[temp['outcome'] == 'Incomplete']
complete.plot(kind='scatter', x='x_start', y='y_start', marker='h', s=50, color='lightsalmon',
alpha=1, ax=ax, zorder=1, label='Complete Pass')
incomplete.plot(kind='scatter', x='x_start', y='y_start', marker='h', s=50, color='slategrey', alpha=1,
ax=ax, zorder=1, label='Incomplete Pass')
for i, event in complete.iterrows():
ax.arrow(event.x_start, event.y_start, event.dx, event.dy, head_width=0.5, ls='-', color='lightsalmon',
alpha=1, zorder=0, length_includes_head=True)
for i, event in incomplete.iterrows():
ax.arrow(event.x_start, event.y_start, event.dx, event.dy, head_width=0.5, ls='-', color='slategrey',
alpha=1, zorder=0, length_includes_head=True)
if event_type == 'Shot':
goals = temp[temp['outcome'] == 'Goal']
misses = temp[temp['outcome'] != 'Goal']
goals.plot(kind='scatter', x='x_start', y='y_start', marker='*', s=75, color='orange',
label='Goal!', alpha=1, ax=ax, zorder=3, lw=0.5, edgecolor='black')
misses.plot(kind='scatter', x='x_start', y='y_start', marker='*', s=75, color='slategrey',
label='Missed Shot', alpha=1, ax=ax, zorder=3, lw=0.5, edgecolor='black')
return fig, ax
def create_heat_map(events: pd.DataFrame, match: str = "an unknown match", fig=None, ax=None,
pitch_length: int = 120, pitch_width: int = 80):
# Draw pitch
fig, ax = create_pitch(pitch_length, pitch_width, fig, ax)
# Filter locations and draw heatmap
location = pd.DataFrame(events.loc[~pd.isnull(events.location)].location.to_list(), columns=['start', 'end'])
sns.kdeplot(location.start, location.end, shade=True, n_levels=10, color="green")
# Set limits
plt.ylim(0, pitch_width)
plt.xlim(0, pitch_length)
plt.title(f"Heatmap of {events.player_name.unique()[0]} during {match}")
return fig, ax
def create_heat_map(events: pd.DataFrame,
match: str = "an unknown match",
fig=None,
ax=None,
pitch_length: int = 120,
pitch_width: int = 80):
"""
This functions creates a heat map for one player
:param events: a data frame containing events from one player
:param match: a string describing the match for which the events are plotted
:param fig: a figure
:param ax: a subplot of the figure on which to plot the arrows
:param pitch_length: a integer describing the length of the pitch
:param pitch_width: a integer describing the width of the pitch
:return fig, ax: a figure showing a heat map
"""
# Draw pitch
fig, ax = create_pitch(pitch_length, pitch_width, fig, ax)
# Filter locations and draw heat map
location = pd.DataFrame(
events.loc[~pd.isnull(events.location)].location.to_list(),
columns=['start', 'end'])
sns.kdeplot(location.start,
location.end,
shade=True,
n_levels=10,
color="green")
# Set limits
plt.ylim(0, pitch_width)
plt.xlim(0, pitch_length)
plt.title(f"Heatmap of {events.player_name.unique()[0]} during {match}")
return fig, ax